Intestinal Bile Secretion Promotes Drug Absorption from Lipid Colloidal Phases via Induction of...

Intestinal Bile Secretion Promotes Drug Absorption from LipidColloidal Phases via Induction of SupersaturationYan Yan Yeap,† Natalie L. Trevaskis,*,† Tim Quach,‡ Patrick Tso,§ William N. Charman,†

and Christopher J. H. Porter*,†

†Drug Delivery, Disposition and Dynamics and ‡Medicinal Chemistry and Drug Action, Monash Institute of Pharmaceutical Sciences,Monash University, 381 Royal Parade, Parkville, Victoria, 3052, Australia§Department of Pathology and Laboratory Medicine, University of Cincinnati, Cincinnati, Ohio 45221, United States

*S Supporting Information

ABSTRACT: The oral bioavailability of poorly water-solubledrugs (PWSD) is often significantly enhanced by coadminis-tration with lipids in food or lipid-based oral formulations.Coadministration with lipids promotes drug solubilization inintestinal mixed micelles and vesicles, however, the mecha-nism(s) by which PWSD are absorbed from these dispersedphases remain poorly understood. Classically, drug absorptionis believed to be a product of the drug concentration in freesolution and the apparent permeability across the absorptivemembrane. Solubilization in colloidal phases such as mixedmicelles increases dissolution rate and total solubilized drugconcentrations, but does not directly enhance (and mayreduce) the free drug concentration. In the absence of changes to cellular permeability (which is often high for lipophilic,PWSD), significant changes to membrane flux are therefore unexpected. Realizing that increases in effective dissolution rate maybe a significant driver of increases in drug absorption for PWSD, we explore here two alternate mechanisms by which membraneflux might also be enhanced: (1) collisional drug absorption where drug is directly transferred from lipid colloidal phases to theabsorptive membrane, and (2) supersaturation-enhanced drug absorption where bile mediated dilution of lipid colloidal phasesleads to a transient increase in supersaturation, thermodynamic activity and absorption. In the current study, collisional uptakemechanisms did not play a significant role in the absorption of a model PWSD, cinnarizine, from lipid colloidal phases. Incontrast, bile-mediated dilution of model intestinal mixed micelles and vesicles led to drug supersaturation. For colloids that wereprincipally micellar, supersaturation was maintained for a period sufficient to promote absorption. In contrast, for primarilyvesicular systems, supersaturation resulted in rapid drug precipitation and no increase in drug absorption. This work suggests thatongoing dilution by bile in the gastrointestinal tract may invoke supersaturation in intestinal colloids and promote absorption,and thus presents a new mechanism by which lipids may enhance the oral absorption of PWSD.

KEYWORDS: absorption, poorly water-soluble drug, lipid-based formulation, supersaturation, bile, micelles, food effect,membrane permeability, cinnarizine

■ INTRODUCTION

The potential for lipid-based formulations (LBF) to enhancethe oral bioavailability of poorly water-soluble drugs (PWSD)has been recognized for over 40 years.1 Lipid coadministrationis thought to enhance the oral absorption of PWSD byproviding mechanisms to overcome both slow dissolution andlow solubility of PWSD in the aqueous gastrointestinal (GI)milieu. First, LBF present drug to the GI tract in a molecularlydispersed form (i.e., in solution in the formulation), therebycircumventing the need for dissolution from the solid to theliquid state. Subsequently, the intercalation of formulationlipids into endogenous lipid digestion pathways results in thegeneration of intestinal lipid colloidal phases (such as vesicularand micellar species) that enhance the solubilization capacity of

the small intestine, promote drug solubilization and reduce therisk of drug precipitation.Solubilization within lipid colloidal phases therefore increases

the apparent solubility of PWSD in the intestinal fluids andpromotes dissolution. However, in the absence of solid drug,solubilization also results in a reduction in thermodynamicactivity.2 In simple micellar systems this reduction inthermodynamic activity is manifest in a decrease in the freeconcentration of drug. Where solubilized drug exists inequilibrium between the free concentration (Cfree) and the

Received: November 15, 2012Revised: March 9, 2013Accepted: March 12, 2013

Article

pubs.acs.org/molecularpharmaceutics

© XXXX American Chemical Society A dx.doi.org/10.1021/mp3006566 | Mol. Pharmaceutics XXXX, XXX, XXX−XXX

pubs.acs.org/molecularpharmaceutics

concentration solubilized in intestinal colloids (Ccolloid), thetotal solubilized drug concentration (Ctotal) is the sum of Cfreeand Ccolloid,

= +C C Ctotal free colloid (1)

Under these circumstances, the solubility of drug in theintermicellar phase (effectively the aqueous solubility of thedrug) provides the upper limit for Cfree and the presence ofsolubilizing colloidal species typically increases the total drugconcentration but does not increase, and often reduces, Cfree.Classical models of passive drug absorption suggest that drugflux across an absorptive membrane is the product of the freedrug concentration and the drug permeability across themembrane. Therefore, where solubilization reduces Cfree (butdoes not alter permeability), absorption is expected to bereduced. Indeed, in solubilized systems, even at saturation, Cfreedoes not exceed the equilibrium solubility of drug in(nonmicellar) aqueous solution. Solubilization in intestinalcolloidal phases therefore provides no practical advantage infree drug concentration and, in the absence of changes topermeability, is unlikely to lead to appreciable increases inmembrane flux when compared to an aqueous solutioncontaining drug at close to saturated solubility. In support ofthis suggestion, many authors have shown that increasing thetotal concentration of a PWSD by solubilization does notnecessarily result in proportional increases in absorptive flux.2−6

Dahan et al. and Miller et al. recently proposed a model toquantify this phenomenon and referred to the existence of a“solubility-permeability interplay” where potential increases inmembrane flux due to increases in solubilized drugconcentration were offset by a reduction in the apparentpermeability, the latter being, in large part, a function ofdecreases in free fraction.5,6

The dispersion and digestion of LBF therefore creates asolubilized reservoir that is in rapid equilibrium with drug infree solution and provides significant advantage in the effectiverate of dissolution of a PWSD when compared to traditionaldissolution from a solid dose form. In contrast, the inherentsolubility limitations to flux (rather than dissolution-ratelimitations) are seemingly unchanged when viewed from theperspective of the free concentration or may be made worse.This appears at odds with the wealth of experimental andpractical observations that suggest the ability of lipids (eitherformulation- or dietary-derived) to enhance the oral absorptionof a range of PWSD.1 A possible explanation for this anomaly isthat the traditional view of drug absorption from colloidaldispersions may not adequately describe the dynamic mannerin which LBF exert absorption-enhancing effects in the GI tract.Two alternative potential mechanisms of drug absorption aretherefore examined here (Figure 1).The first mechanism evaluated was the potential for drug

absorption to occur via direct collisional transfer from lipidcolloidal phases to the absorptive membrane, and thus to bemediated not only by Cfree but also by the solubilized fraction,Ccolloid. Previously, studies by Storch and colleagues have shownthat the transfer of poorly water-soluble fatty acids betweenmodel cell membranes and proteins may occur via collisionaltransfer.7,8 More recently, the possibility of drug absorption viacollisional uptake has been suggested by Yano et al. and Gao etal.9,10 Collisional uptake may or may not be receptor-mediated,11 however, lipid uptake receptors such as CD36,12

SR-BI13 and NPC1L114 have been suggested to facilitate theabsorption of cholesterol and fatty acids, poorly water-soluble

molecules that are also solubilized in intestinal colloidal phases.In the case of SR-BI and CD36, direct interaction of thereceptor with colloidal structures such as HDL (high-densitylipoprotein), bile salt micelles and phospholipid vesicles13,15 hasalso been suggested, raising the possibility that lipid uptakereceptors may interact directly with intestinal lipid colloidalphases to facilitate collisional absorption of solubilized contents,including PWSD.The second mechanism evaluated was the potential for

endogenous lipid processing pathways to lead to drugsupersaturation in lipid colloidal phases. Supersaturationincreases the thermodynamic activity of solubilized drug and,in the solubilization model described by eq 1, will increase Cfreeabove the equilibrium aqueous drug solubility. This in turn isexpected to enhance drug flux. The potential for super-saturation to enhance the oral bioavailability of PWSD hasreceived increasing recent interest.10,16 For LBF, supersatura-tion may be generated by the loss of drug solubilizationcapacity resulting from the digestion of triglycerides17−19 and/or surfactants,20 and the dilution of cosolvents21 during GIprocessing.19 In contrast, the possibility that supersaturationmay result directly from interactions between lipid colloidalphases and biliary fluids has been almost entirely ignored.Although traditional micellar solubilization models suggest thatincreases in bile salt concentrations increase drug solubilization,previous studies have also shown that dilution of lipid colloidalphases with model intestinal fluids (containing bile salts,phospholipid and cholesterol) may lead to the generation ofless lipid-rich colloidal phases with lowered solubilizationcapacities.22,23 This provides a plausible mechanism forsupersaturation generation in the small intestine and has beenexamined in detail here.

Figure 1. Alternative mechanisms of drug absorption from intestinallipid colloidal phases. In collisional drug absorption (left panel), inaddition to the diffusion of free drug molecules (a), lipid colloidalphases collide with the absorptive membrane and facilitate directtransfer of solubilized drug into absorptive cells (b). Collisionaltransfer may or may not be receptor-mediated. In supersaturation-enhanced drug absorption (right panel), the interaction betweensecreted bile and lipid colloidal phases leads to drug supersaturation(possibly via stimulation of phase changes to less lipid-rich colloidswith lowered drug solubilization capacity). The increase inthermodynamic activity manifests in increases in free drugconcentration, and enhanced diffusional flux (a) across the absorptivemembrane.

Molecular Pharmaceutics Article

dx.doi.org/10.1021/mp3006566 | Mol. Pharmaceutics XXXX, XXX, XXX−XXXB

The data suggest, that under the conditions explored,collisional drug absorption has little impact on drug absorptionfrom intestinal colloidal species. In contrast, supersaturation-enhanced drug absorption, mediated by the interaction betweensecreted bile and lipid colloidal phases, may provide anendogenous mechanism to promote supersaturation and tofacilitate drug absorption from lipid colloidal phases.

■ EXPERIMENTAL SECTIONMaterials. Cinnarizine, flunarizine dihydrochloride, mon-

ensin sodium, sodium taurocholate, sodium taurodeoxycholate,sodium glycocholate, sodium glycochenodeoxycholate, choles-terol, L-α-lysophosphatidylcholine (LPC, from egg yolk), L-α-phosphatidylcholine (PC, from dried egg yolk), oleic acid,caprylic acid, monocaprylin, N-hydroxysulfosuccinimide so-dium, dicyclohexylcarbodiimide solution (60% w/v in xylenes),N,N-dimethylformamide, Tween 80, potassium dihydrogenphosphate (KH2PO4) and sodium chloride (NaCl) wereobtained from Sigma-Aldrich, Australia. Sodium taurocheno-deoxycholate, sodium glycodeoxycholate, ortho-phosphoricacid 85% (H3PO4), sodium hydroxide pellets (NaOH), tert-butyl methyl ether (TBME), dimethyl sulfoxide (DMSO),glacial acetic acid and absolute ethanol were from Merck,Australia. Disodium hydrogen orthophosphate (Na2HPO4),sodium dihydrogen orthophosphate (NaH2PO4·2H2O) andammonium dihydrogen orthophosphate (NH4H2PO4) (AjaxFinechem, Australia), cholesterol, [4-14C] (49.8 mCi/mmol),and Irga-Safe Plus (Perkin-Elmer Life Sciences, Waltham, MA),oleic acid, [9,10-3H(N)] (60 Ci/mmol) (American Radio-labeled Chemicals, St. Louis, MO), Block Lipid Transport-1(BLT-1, i.e., 2-hexyl-1-cyclopentanone thiosemicarbazone)(Chembridge, San Diego, CA), ezetimibe (Jai Radhe Sales,AMD, India), heparin sodium injection BP (1000 IU/mL,Hospira, Australia), xylazine (100 mg/mL, Troy Laboratories,Australia), acepromazine (10 mg/mL, Ceva Delvet, Australia),ketamine (100 mg/mL, Provet, Australia) and pentobarbitonesodium (325 mg/mL, Virbac, Australia) were obtained fromlisted suppliers. Acetonitrile, methanol and chloroform wereanalytical reagent grade. Water was obtained from a MilliporeMilli-Q Gradient A10 water purification system (Millipore,Billerica, MA).Sulfo-N-succinimidyl Oleate (SSO) Synthesis. SSO syn-

thesis was adapted from the procedure of Harmon et al.24

Briefly, dicyclohexylcarbodiimide (1.26 mmol) and N-hydrox-ysulfosuccinimide (sodium salt, 1.20 mmol) were added to asolution of oleic acid (1.20 mmol) dissolved in N,N-dimethylformamide (2 mL), and the reaction was stirred atroom temperature overnight. The precipitated dicyclohexylureawas removed by filtration and ethyl acetate (2 mL) added tothe filtrate, which was left to stand at 4 °C overnight. SSO(precipitate) was then collected by filtration and dried undervacuum (1 mmHg). SSO identity was confirmed by NMR andmass spectrometry.Experimental outline. To assess the role of receptor-

mediated collisional drug absorption, cinnarizine bioavailabilitywas assessed after administration of a lipid emulsionformulation in the absence and presence of BLT-1, SSO andezetimibe. BLT-1, SSO and ezetimibe are chemical inhibitors ofSR-BI,25 CD3626 and NPC1L1,27 respectively. The possibilityof endocytosis-mediated uptake was also investigated by the useof a general endocytosis inhibitor, monensin.The role of collisional drug absorption was assessed more

generically using an in situ rat jejunum perfusion model to

compare the absorptive flux of cinnarizine from two distinctlydifferent lipid colloidal phases (micelles vs vesicles) withmatched drug solubilization capacities. Colloidal systems withthe same total solubilization capacity, loaded with drug at thesame concentration, have the same thermodynamic activity andtherefore Cfree is the same in both cases. Under thesecircumstances, comparison of the flux profiles obtained fromtwo structurally different colloids, but with identical Cfree,provides a means of determining whether the nature of thecolloid, or Cfree, is the principal determinant of absorption.Generation of identical flux profiles from both systems wouldtherefore confirm the dependence of flux on thermodynamicactivity and free concentration, whereas a significant differencein flux would indicate a role for factors beyond Cfree indetermining flux. These include the potential for collisionaldrug absorption mechanisms since collision rates are astatistical function of particle number and are expected to bemarkedly higher for micelles (where the smaller particle sizeresults in higher particle numbers) when compared to vesicles.To assess the potential for intestinal fluids to enhance drug

absorption from lipid colloidal phases via the induction of drugsupersaturation, whole bile was collected from fasted rats andmixed with model micelles and vesicles to simulate the processof interaction with bile in vivo. The potential for bile to generatedrug supersaturation was evaluated in vitro by assessing changesin cinnarizine solubility, and by monitoring the kinetics ofcinnarizine solubilization and precipitation, following bileaddition to cinnarizine-loaded micelles and vesicles. Sub-sequently, the impact of drug supersaturation on the intestinalabsorptive flux of cinnarizine from micelles and vesicles wasassessed in an in situ rat jejunum perfusion model, with andwithout coperfusion of donor bile. Finally, the relevance of bile-induced drug supersaturation in vivo was assessed viaexamination of changes to cinnarizine bioavailability afteradministration of drug-loaded micelles and vesicles (withmatched thermodynamic activity) in bile-intact vs bile-divertedrats.

Formulation Preparation. Lipid Emulsion. The lipidemulsion (3 mL per dose) consisted of 1 mg of cinnarizineand 49 mg of oleic acid solubilized in 8 mM sodiumtaurocholate, 2 mM phosphatidylcholine, 2 mM cholesteroland trace amounts of 14C-cholesterol (1 μCi/3 mL) and/or 3H-oleic acid (3 μCi/3 mL). The emulsion was prepared in 7.5 mLbatches by weighing appropriate masses of cinnarizine in oleicacid stock solution (20 mg/g), phosphatidylcholine andcholesterol into a glass vial, and the mixture made up tovolume with a buffered sodium taurocholate solution (bufferconsisted 18 mM Na2H2PO4·2H2O and 12 mM Na2HPO4).Appropriate volumes of 14C-cholesterol, 3H-oleic acid, 5 mg/mL BLT-1 in ethanol (for SR-BI inhibition experiments only),25 mg/mL SSO in DMSO (for CD36 inhibition experimentsonly) and 10 mg/mL monensin in ethanol (for endocytosisinhibition experiments only) were spiked into the vial andvortexed for 1 min. The formulation was emulsified byultrasonification with a Misonix XL 2020 ultrasonic processor(Misonix, Farmingdale, NY) equipped with a 3.2 mmmicroprobe tip running at an amplitude of 240 μm and afrequency of 20 kHz for 1.5 min. The total solventconcentration in the emulsion was ≤2.5% v/v. The emulsionwas used within 4 h of preparation, and the concentration ofdrug and labeled cholesterol and/or oleic acid was assayedbefore dosing (in duplicate) to confirm compound content inthe emulsion and to allow for dose normalization between rats.


dx.doi.org/10.1021/mp3006566 | Mol. Pharmaceutics XXXX, XXX, XXX−XXXC

Intravenous Formulation. The formulation (1 mL perdose) used for intravenous administration of cinnarizine and14C-cholesterol comprised 0.5% w/v Tween 80 in buffer (36mM Na2HPO4 and 22 mM KH2PO4, adjusted to pH 4 withacetic acid). Cinnarizine and cholesterol were added to theformulation by spiking 5% v/v DMSO (containing 10 mg/mLcinnarizine) and 5% v/v ethanol (containing 5 mg/mLcholesterol and 40 μCi/mL 14C-cholesterol) into the micellarsolution. The formulation was mixed by vortexing, and theconcentration of drug and labeled cholesterol was assayedbefore dosing to confirm compound content in the formulationand to allow for dose normalization between rats. Theformulation was used within 1 h of preparation.Model Micelles and Vesicles. The preparation of model

micelles and vesicles was guided by the methods and phasediagram published by Kossena et al.23 Medium-chain lipidcontaining colloids were chosen over long-chain systems sincethe former have previously been shown to generate mono-phasic micellar and vesicular systems.23 Relatively high lipidconcentrations were chosen to reflect the species that areexpected to initially form during the digestion of medium-chaintriglycerides.18 The model colloids consisted of tricaprylindigestion products (caprylic acid and monocaprylin) solubilizedin simulated endogenous intestinal fluid (SEIF). SEIFcomprised the six most prevalent bile salts in human bile,28

lysophosphatidylcholine (LPC) and cholesterol. The total bilesalt:LPC:cholesterol molar ratio was maintained at 16:4:1,reflecting known ratios within fasted human bile.29,30 Thecombination of bile salts used here comprised 25 mol % sodiumglycocholate, 17.5 mol % sodium glycodeoxycholate, 25 mol %sodium glycochenodeoxycholate, 12.5 mol % sodium taur-ocholate, 7.5 mol % sodium taurodeoxycholate and 12.5 mol %sodium taurochenodeoxycholate. The concentration ratios ofthe bile salts were chosen based on average concentrations ofthe six most prevalent bile salts found in human bile.28 Thecaprylic acid:monocaprylin molar ratio was kept at 2:1,reflecting the ratio of digestion products expected on digestionof 1 mol of triglyceride. The concentration of micellar andvesicular components was varied by trial and error (butmaintaining the ratios described above) to identify systems withsimilar drug solubilization capacities. It has previously beenshown that the thermodynamic activity (i.e., free concen-tration) of drug in a solubilized system may be estimated viaassessment of solubility behavior, such that having differentcolloidal solutions containing drug at a fixed proportion of thesaturated solubility results in matched free concentrations.31

Thus, drug was loaded into either micellar or vesicular systemsat the same concentration (and the same proportion ofsaturated solubility) and was therefore present at the samethermodynamic activity (i.e., Ctotal, Ccolloid and Cfree were thesame in both micellar and vesicular systems). The compositionsof the identified micellar and vesicular systems are shown inTable 1.SEIF (8 mM total bile salt:2 mM LPC:0.5 mM cholesterol)

was prepared in 50 mL batches. Briefly, LPC and cholesterolwere dissolved in 1 mL of chloroform in a round-bottom flask,followed by solvent evaporation under vacuum. The thin filmformed by solvent evaporation was reconstituted with bufferedbile salt solution (8 mM total bile salt, 18 mM NaH2PO4·2H2Oand 12 mM Na2HPO4, 100 mM NaCl), vortexed for 1 min andallowed to equilibrate at room temperature overnight. Whenvesicles were prepared, a similar procedure was adopted, but inthis case SEIF was diluted 4-fold with buffer (18 mM

NaH2PO4·2H2O and 12 mM Na2HPO4, 108 mM NaCl) toreduce the bile salt:lipid concentration ratio. Micelles andvesicles were prepared in 10 mL batches by adding caprylic acidand monocaprylin (quantities in Table 1) to SEIF, followed bypH adjustment to 6.30 with solid NaOH and vortexing for 1min. The phases were then ultrasonicated as described earlier(30 s continuous ultrasonication followed by pulsatile, 1 s on/1s off ultrasonication for 5 min). When included in the colloids,cinnarizine was predissolved in caprylic acid, and the drug/fattyacid solution was allowed to equilibrate overnight prior tomicelle/vesicle preparation.

Particle Sizing. The particle size of model micelles andvesicles was determined by photon correlation spectroscopy(Malvern Instruments Nano-ZS Zetasizer, Malvern, U.K.).Micelles had a mean particle size of 9 ± 1 nm [polydispersityindex 0.128 ± 0.040], and vesicles had a mean particle size of443 ± 32 nm [polydispersity index 0.456 ± 0.038]. Datareported are mean ± SEM of n = 3 determinations.

Equilibrium Solubility of Cinnarizine in the ModelMicelles and Vesicles. Excess solid cinnarizine was added to2 mL micelles or vesicles in glass vials. Vials were brieflyvortexed and incubated at 37 °C and samples taken every 24 hover a period of 120 h. During sampling, vials were centrifuged(2200g, 10 min, 37 °C), 50 μL of supernatant was sampled, andvials were revortexed. Equilibrium solubility was defined whendrug concentrations in consecutive samples varied by ≤5% andwas determined on three separate occasions.Equilibrium solubility of cinnarizine was also determined

after 1:1 v/v addition of bile/bile pH 6.30/buffer pH 6.30 tothe different colloidal phases. The pH of fresh bile was adjustedto 6.30 with H3PO4. Buffer pH 6.30 consisted of 18 mMNaH2PO4·2H2O, 12 mM Na2HPO4 and 108 mM NaCl.

Kinetics of Cinnarizine Precipitation after theAddition of Bile to Model Micelles and Vesicles. Thekinetics of cinnarizine precipitation was monitored afteraddition of bile to the micellar and vesicular phases, todetermine whether a period of drug supersaturation existedprior to drug precipitation. In a temperature and stirring rate-controlled vessel, 2.5 mL of bile was added to 2.5 mL ofmicelles or vesicles containing 0.2 mg/mL cinnarizine (∼80%saturated solubility). Samples (100 μL) were taken before theaddition of bile, and at 1, 10, 20, 30, 40, 50, 60, 80, 100, 120min after bile addition. Samples were immediately centrifuged(2200g, 5 min, 37 °C) to separate precipitated drug, and 50 μLof supernatant was assayed for drug content. The proportion ofthe initial solubilized cinnarizine concentration that remainedsolubilized after bile addition was assessed as the percent of thedrug mass remaining in solution (i.e., concentration in the

Table 1. Composition of Model Micelles and Vesiclesa

concentration (mM)

total bilesaltb LPCc cholesterol

caprylicacid monocaprylin

micelles 8 2 0.5 69.3 34.7vesicles 2 0.5 0.125 52.0 26.0

aMicelles and vesicles also consist of 18 mM NaH2PO4·2H2O and 12mM Na2HPO4. Sodium strength was adjusted to 150 mM with NaCl.Final pH of phases was adjusted to 6.30 ± 0.01. bTotal bile salt consistof 25 mol % sodium glycocholate, 17.5 mol % sodium glycodeox-ycholate, 25 mol % sodium glycochenodeoxycholate, 12.5 mol %sodium taurocholate, 7.5 mol % sodium taurodeoxycholate, 12.5 mol% sodium taurochenodeoxycholate. cLPC is lysophosphatidylcholine.


dx.doi.org/10.1021/mp3006566 | Mol. Pharmaceutics XXXX, XXX, XXX−XXXD

supernatant multiplied by the volume remaining in vessel)relative to the total drug mass in the vessel at each time point.Solid-State Analysis of the Cinnarizine Precipitate

Using Polarized Light Microscopy. Selected cinnarizinepellets from the precipitation kinetics experiments wereanalyzed using a Zeiss Axiolab microscope (Carl Zeiss,Oberkochen, Germany) equipped with crossed polarizingfilters. At the end of the precipitation kinetics experiments,1.5 mL of remaining bile + colloid mixture was centrifuged(2200g, 10 min, 37 °C), the supernatant was discarded, and asmall amount of pellet was carefully placed on a microscopeslide. Samples were analyzed under cross-polarized light, andimages were recorded using a Canon PowerShot A70 digitalcamera (Canon, Tokyo, Japan).Animals. All rat studies were approved by the institutional

animal ethics committee and were conducted in accordancewith the guidelines of the Australian and New Zealand Councilfor the Care of Animals in Research and Teaching. MaleSprague−Dawley rats (280−330 g) were used in all experi-ments and were allowed to acclimatize in the institutionalanimal housing facility for at least 7 days with free access tostandard chow and water. All animals were fasted overnight(12−18 h) prior to surgery.Surgical Procedures. Anesthesia was induced in rats by

subcutaneous injection of 1.0 mL/kg of “Cocktail I” (37.3 mg/mL ketamine, 9.8 mg/mL xylazine, 0.4 mg/mL acepromazinein saline), and maintained throughout the study withsubcutaneous doses of 0.44 mL/kg of “Cocktail II” (90.9mg/mL ketamine, 0.9 mg/mL acepromazine) when required.Rats were maintained on a 37 °C heated pad throughoutsurgery and experiments. At the end of all experiments, ratswere euthanized via an intravenous or intracardiac injection of100 mg of sodium pentobarbitone.Cinnarizine Bioavailability Studies following Intraduode-

nal Administration. The surgical procedures for the conduct ofbioavailability studies included cannulations of the right carotidartery, right jugular vein, duodenum (1 cm below pylorus) andcommon bile duct (only for bile-diverted rats). The surgicalprocedures for the cannulations were as described previ-ously.32,33

Fasted Rat Bile Collection. The bile duct was cannulatednear the hilum of the liver (where the duct is free of pancreatictissue) in order to facilitate the collection of bile fluid withoutcontamination by exocrine pancreatic secretions.34 Rats wererehydrated via saline infusion (1.5 mL/h) into a cannulainserted into the right jugular vein, and bile was continuouslycollected for 5 h. The concentration of total bile salt incollected bile was assayed using a validated enzymaticcolorimetric assay (Total Bile Acids kit #431-15001; WakoPure Chemical Industries, Osaka, Japan) on a plate reader(Fluostar Optima plate reader, BMG Labtechnologies,Germany) measuring absorbance at a wavelength of 540 nm.In all subsequent experiments, bile was used within 24 h ofcollection.Single-Pass Rat Jejunum Perfusion. The model employed

to assess flux across rat jejunum involved in situ perfusion(single-pass) of an isolated jejunal segment and simultaneousblood collection from the corresponding mesenteric veinbranch. The surgical procedures for the perfusion studies aresimilar to those described elsewhere, with slight modifica-tions.35 Briefly, the right jugular vein was cannulated to enableinfusion of donor blood. A piece of jejunal segment (∼10 cm2)was isolated and cannulated at the proximal and distal ends

with sections of Teflon tubing (0.03 in. i.d. proximal/inlet,Upchurch Scientific, Oak Harbor, WA; 0.0625 in. i.d. distal/outlet, Shimadzu, Kyoto, Japan). Jejunal contents were initiallyflushed with warm perfusion buffer (150 mM Na+, 18 mMH2PO4

−, 12 mM HPO42−, 108 mM Cl−, adjusted to pH 6.30 ±0.01). The mesenteric vein draining the jejunal segment wasthen isolated, the rat heparinized (90 IU/kg) via the jugularvein and the mesenteric vein immediately catheterized. A dropof superglue was placed over the site of catheterization, andsilicone tubing (0.025 in. i.d., Helix Medical, CA) attached tothe catheter tip for the collection of venous blood. Immediatelyfollowing catheterization of the mesenteric vein and for theremainder of the experiment, rats were infused with heparinizeddonor rat blood (5 IU/mL) via the jugular vein, and the rate(0.3 mL/min) was adjusted based on the outflow from themesenteric blood.

Cinnarizine Bioavailability after Intraduodenal Infu-sion in Anesthetized Rats. A 30 min equilibration periodwas allowed between the end of surgery and drug dosing. Toexamine the impact of lipid uptake receptors on drugabsorption, studies were conducted in the presence or absenceof lipid uptake inhibitors. To examine the impact of bile-induced drug supersaturation on drug absorption, studies wereconducted in bile-intact or bile-diverted rats. Colloidal systems(lipid emulsion, micelles, vesicles) containing cinnarizine wereinfused into the duodenum of rats at a rate of 1.5 mL/h for 2 h.When the infusion was complete, saline was infused at a rate of1.5 mL/h for 10 min to flush any remaining formulation in thetubing into the duodenum. Blood samples (0.3 mL) werecollected via the carotid artery cannula up to 8 h after infusioninitiation into tubes containing 3 IU heparin. The samplingintervals were as follows: t = 0, 1, 2, 2.5, 3, 4, 6, 8 h for thereceptor inhibition studies; and t = 0, 1, 1.5, 2, 3, 4, 6, 8 h forthe bile-induced drug supersaturation studies. After each bloodsample was taken, the cannula was flushed with 0.3 mL of 2 IU/mL heparinized saline to ensure cannula patency, and to replacethe volume of blood removed. Plasma was separated bycentrifugation (10000g, 5 min) to enable analysis of drug andlabeled lipid content.

Administration of Lipid Uptake Inhibitors (BLT-1, SSO,Ezetimibe) and Endocytosis Inhibitor (Monensin). Ezetimibehas previously been dosed at 0.3 mg/kg intravenously into rats,and has been shown to inhibit cholesterol absorption (from anintraduodenally dosed lipid emulsion) without reports oftoxicity.36 Therefore, in our study, intravenous administrationof ezetimibe (0.3 mg/kg via the jugular vein) was selected asthe route to administer the inhibitor at the beginning of the 30min equilibration period. An appropriate volume of 5 mg/mLezetimibe in ethanol was spiked into blank rat plasma, and 0.8mL of resultant plasma was dosed into rats as an intravenousbolus. The total ethanol concentration was less than 2.5% v/v.In the case of BLT-1 and SSO, the inhibitors had not beenpreviously administered intravenously, therefore local (i.e.,intestinal) administration was selected to limit the systemiceffects of the inhibitors. Monensin was also administereddirectly into the intestine to minimize systemic endocytosisinhibition. Here, 100 μM BLT-1, 1 mM SSO and 100 μMmonensin were preinfused intraduodenally (in saline) at a rateof 1.5 mL/h during the 30 min equilibration period, andsubsequently coinfused at the same concentration as part of thelipid emulsion. BLT-1 and SSO have previously been shown toinhibit lipid uptake in cell-based studies at concentrations of 1−


dx.doi.org/10.1021/mp3006566 | Mol. Pharmaceutics XXXX, XXX, XXX−XXXE

10 μM25 and 400 μM,24 respectively. Monensin has previouslybeen shown to inhibit endocytosis in cultured cells at 10 μM.37

Intravenous administration studies were also conducted incontrol rats and monensin-treated rats to assess the effect ofendocytosis inhibition on the systemic distribution andclearance of cinnarizine and cholesterol. In these studies,blank (i.e., not containing cinnarizine and 14C-cholesterol) lipidemulsion (with or without 100 μM monensin) was infusedintraduodenally as described above, and intravenous infusion ofthe cinnarizine and 14C-cholesterol containing intravenous (iv)formulation was commenced at the same time. The ivformulation (1 mL) was infused into the right jugular vein ata rate of 0.05 mL/15 s (total infusion period <5 min). Bloodsamples (0.3 mL) were collected via the carotid artery cannulaat t = 5, 15, 30, 60, 120, 180, 240, 360, 480 min after infusioninitiation into tubes containing 3 IU of heparin. After eachblood sample was taken, the cannula was flushed with 0.3 mL of2 IU/mL heparinized saline. Plasma was separated bycentrifugation (10000g, 5 min) to enable analysis of drug andlabeled cholesterol content.Intestinal Absorptive Flux Assessment via in Situ

Single-Pass Rat Jejunum Perfusion. After surgery, animalswere equilibrated for 30 min, during which time heparinizeddonor rat blood was infused via the jugular vein as describedabove. During re-equilibration, blood from the cannulatedmesenteric vein (∼0.3 mL/min) was collected for reinfusion.Perfusion buffer was pumped through the jejunal segment at arate of 0.1 mL/min and outflowing buffer discarded to waste.The exposed jejunal segment was kept moist by covering withsaline-soaked gauze throughout the experiment.In all experiments, the concentration of cinnarizine in the

perfusate was held at 0.1 mg/mL (∼40% saturated solubility).Therefore, in experiments where micelles or vesicles wereperfused alone, cinnarizine was loaded into the perfusate at 0.1mg/mL. In experiments where micelles or vesicles werecoperfused in a 1:1 v/v ratio with a secondary perfusate ofbile/bile pH 6.30/buffer pH 6.30, cinnarizine was loaded intothe primary perfusate at 0.2 mg/mL, such that 1:1 v/v dilutionled to a final perfusate concentration of 0.1 mg/mL.Perfusate flow was maintained at 0.1 mL/min in all

experiments to minimize variations in the thickness of theunstirred water layer.38 For experiments where 1:1 v/vcoperfusion of the phases with bile/bile pH 6.30/buffer pH6.30 was required, micelles/vesicles and bile/buffer werepumped at 0.05 mL/min, and mixed via a three-way “T”connector immediately prior to entry into the jejunal segment,providing a total perfusate flow of 0.1 mL/min. Perfusate wassampled at t = 0 to confirm lipid and drug concentrations. Afterthis time, the outflowing perfusate was continuously collectedat 10 min intervals and briefly vortexed before samples weretaken for analysis of drug and lipid content. For experimentswhere drug supersaturation was generated, perfusate sampleswere taken before and after centrifugation (2200g, 2 min), toassess the degree of drug precipitation within the jejunalsegment. Blood draining the perfused jejunal segment wascollected at 5 min intervals, plasma was separated bycentrifugation (10000g, 5 min) and samples were taken foranalysis of drug content by HPLC as described below.Analytical Procedures. Sample Preparation and HPLC

Assay Conditions for Cinnarizine. Samples of lipid emulsionwere prepared for HPLC assay by an initial 80-fold dilutionwith chloroform:methanol (2:1 v/v), followed by a 10-folddilution with mobile phase (50% v/v acetonitrile:50% v/v 20

mM NH4H2PO4). Samples of iv formulation and micelles/vesicles were prepared for HPLC assay by a 900-fold and a 400-fold dilution with mobile phase (50% v/v acetonitrile:50% v/v20 mM NH4H2PO4), respectively. Plasma samples wereprepared for HPLC using a validated extraction procedure,with flunarizine as an internal standard, as reportedpreviously.39

Cinnarizine HPLC assay conditions were as describedpreviously,39 with slight modification to the mobile phaseemployed for cinnarizine quantification in plasma, to 45% v/vacetonitrile:55% v/v 20 mM NH4H2PO4 in this study. Replicateanalysis of n = 4 quality control samples revealed acceptableaccuracy and precision (±10%, ±15% at the limit ofquantification) for concentrations between 20 and 1000 ng/mL for iv formulation, emulsion, micelles and vesicles, and 10and 320 ng/mL for plasma.

Scintillation Counting. Quantification of 14C-cholesteroland 3H-oleic acid in the plasma was performed via scintillationcounting on a Packard Tri-Carb 2000CA liquid scintillationanalyzer (Packard, Meriden, CT, USA). Plasma samples (50μL) were added to 2 mL of Irga-safe Plus scintillation fluidfollowed by a 10 s vortex. Samples were corrected forbackground radioactivity by the inclusion of a blank plasmasample in each run. The counting method was validated byspiking blank plasma with low, medium and high concen-trations of labeled cholesterol and oleic acid (in triplicate). Themeasured concentrations were within 5% of the nominalconcentration.

Blood:Plasma Ratio Determination for Cinnarizine. Theblood:plasma ratio for cinnarizine was determined by spiking0.5 mL of blank blood with known amounts of cinnarizine toachieve low, medium and high concentrations (in triplicate).Plasma was separated by centrifugation (10000g, 5 min) andplasma drug concentration assayed by HPLC. The blood:-plasma ratio was calculated from the ratio of knownconcentration in spiked blood to the concentration measuredin plasma separated from spiked blood. The mean blood:-plasma ratio was subsequently used to convert plasmaconcentrations to blood concentrations in perfusion experi-ments, enabling quantification of total drug transport intomesenteric blood.

Calculations. In the single-pass rat jejunum perfusion model,permeability coefficients were calculated using steady-state drugconcentrations in perfusate and blood. Two apparentpermeability coefficients (Papp) were calculated as describedpreviously:35

= − ·PCC

disappearanceQA

lnapp1

0 (2)

=·⟨ ⟩

ΔΔ( )

PA C

appearance

Mt

app

B

(3)

where “disappearance Papp” is the apparent permeabilitycoefficient calculated from drug loss from the perfusate (cm/s); “appearance Papp” is the apparent permeability coefficientcalculated from drug appearance in the mesenteric blood (cm/s); Q is the perfusate flow rate (mL/s); A is the surface area ofthe perfused jejunal segment (cm2), which is calculated bymultiplying the diameter by the length of the perfused intestinalsegment as described previously;40 C1 is the average steady-state drug concentration exiting the perfused jejunal segment(ng/mL); C0 is the drug concentration entering the jejunal


dx.doi.org/10.1021/mp3006566 | Mol. Pharmaceutics XXXX, XXX, XXX−XXXF

segment (ng/mL); ΔMB/Δt is the average rate of drug massappearance in mesenteric blood at steady state (ng/s); and ⟨C⟩is the logarithmic mean drug concentration in the lumen (ng/mL), where ⟨C⟩ = (C1 − C0)/(ln C1 − ln C0).Noncompartmental Pharmacokinetic Analysis. The max-

imum plasma concentration (Cmax), time to reach Cmax (Tmax),area under the plasma concentration−time curve from timezero to the last measured concentration (AUC0−8h), area underthe plasma concentration−time curve extrapolated to infinity(AUC0−inf), elimination rate constant (Ke), volume ofdistribution (Vd) and clearance (Cl) were calculated usingWINONLIN version 5.3 (Pharsight Inc., Apex, NC, USA).Statistical Analysis. Results were analyzed using Student’s

t test. A P value of <0.05 was considered to be a significantdifference. Analyses were performed using SPSS v19 forWindows (SPSS Inc., Chicago, IL, USA).

■ RESULTS

SR-BI, CD36, NPC1L1 and Endocytosis Have LittleImpact on Drug Absorption from Intestinal ColloidalPhases. Inhibition of the lipid uptake receptors SR-BI (bycoinfusion of 100 μM BLT-1), CD36 (by coinfusion of 1 mMSSO) and NPC1L1 (by intravenous administration of 0.3 mg/kg ezetimibe) did not result in significant changes to thesystemic plasma concentration−time profiles or systemicexposure of cinnarizine (Figure 2A, Table 2). Inhibition ofendocytosis (by coinfusion of 100 μM monensin) unexpectedly

led to significant increases in cinnarizine systemic plasmaconcentration at t = 2.5 and 4 h (Figure 2A), and increased theAUC0−8h 1.5-fold (Table 2). However, subsequent intravenousdosing studies revealed significantly lower Vd and Cl values inthe monensin-treated rats when compared to control (Table S1in the Supporting Information), suggesting that the increase incinnarizine systemic exposure was due to a decrease incinnarizine systemic distribution and clearance, rather thanchanges to intestinal absorption. The data suggest a limited rolefor SR-BI, CD36, NPC1L1 and endocytosis generally, in theabsorption of cinnarizine from a lipid emulsion.In contrast, inhibition of NPC1L1 did lead to significantly

lower systemic plasma concentrations of exogenously dosedcholesterol at t = 2, 2.5, 3, 4, 6, 8 h (Figure 2B). Plasmaconcentrations of exogenously dosed oleic acid also appearedlower in rats following inhibition of CD36 although differenceswere not significantly different (Figure 2C). While inhibition ofSR-BI and CD36 did not lead to significant changes to thesystemic plasma concentration−time profile of exogenouslydosed cholesterol, inhibition of endocytosis significantlyincreased the systemic plasma concentrations of exogenouslydosed cholesterol at t = 4, 8 h (Figure 2B). This increase inexposure is likely explained by a decrease in systemicdistribution of 14C-cholesterol in the monensin-treated rats,as plasma 14C-cholesterol concentrations at early sample timepoints (which reflect the distribution phase) in intravenousstudies were significantly higher in the monensin-treated rats(Figure S1B in the Supporting Information) when compared tocontrols (Vd and Cl could not be calculated for cholesterol as atypical elimination phase was not evident in the systemicplasma concentration vs time profiles of cholesterol: Figure 2Band Figure S1B in the Supporting Information).

Drug Absorption from Micelles and Vesicles IsDetermined by Cfree and Not Colloidal Structure. Theintestinal perfusion of colloidal media with markedly differentcompositional profiles (micelles and vesicles) but withcomparable Cfree and thermodynamic activity did not result insignificant differences in steady-state absorptive flux, disappear-ance Papp or appearance Papp of cinnarizine (Table 3, Figure 3A:filled symbols). Similarly, in spite of large differences in particlesize and composition, the intraduodenal infusion of the samemicellar or vesicular systems to bile-diverted rats did not resultin significant differences in systemic plasma concentration−time profiles (Figure 4: filled symbols) and pharmacokineticparameters of cinnarizine (Table 5: micelles vs vesicles in bile-

Figure 2. Systemic plasma concentration−time profiles of (A)cinnarizine (CIN), (B) 14C-labeled cholesterol (Ch) and (C) 3H-labeled oleic acid (OA) following intraduodenal infusion of a 3 mLlipid emulsion consisting of 1 mg of cinnarizine emulsified in 59 mMoleic acid, 8 mM sodium taurocholate, 2 mM phosphatidylcholine, 2mM cholesterol, 1 μCi of 14C-cholesterol and/or 3 μCi of 3H-oleicacid. Experiments were performed in control rats (filled circle); ratstreated with 100 μM BLT-1 (open circle); rats treated with 1 mM SSO(open triangle); rats treated with 0.3 mg/kg ezetimibe (open square);and rats treated with 100 μM monensin (open diamond), to inhibitthe lipid uptake receptors SR-BI, CD36, NPC1L1 and endocytosis,respectively. BLT-1, SSO, and monensin were coinfused as part of thelipid emulsion; ezetimibe was administered intravenously. Datarepresent mean ± SEM of n = 4 rats for A and B; and n = 3 ratsfor C. Statistically significant difference with respect to control rats (p< 0.05) is denoted by the symbol *.

Table 2. Pharmacokinetic Parameters for Cinnarizine afterIntraduodenal Administration of a 3 mL Lipid Emulsion inRatsa

experimental group AUC0−8h (ng h/mL) Cmax (ng/mL) Tmax (h)

control 978 ± 122 288.1 ± 26.6 2.4 ± 0.1BLT-1-treated 883 ± 21 219.0 ± 20.0 2.6 ± 0.5SSO-treated 924 ± 88 195.5 ± 33.6 2.4 ± 0.2ezetimibe-treated 1048 ± 92 333.9 ± 13.6 2.3 ± 0.1monensin-treated 1487 ± 103b 432.8 ± 45.9b 2.3 ± 0.1

aIn each case, the formulation type was emulsion and the CIN dosewas 3.33 mg/kg. Experiments were performed in control rats; ratstreated with 100 μM BLT-1; rats treated with 1 mM SSO; rats treatedwith 0.3 mg/kg ezetimibe; and rats treated with 100 μM monensin, toinhibit the lipid uptake receptors SR-BI, CD36, NPC1L1, andendocytosis, respectively. Values represent mean ± SEM of n = 4rats. bSignificant difference when compared to control group.


dx.doi.org/10.1021/mp3006566 | Mol. Pharmaceutics XXXX, XXX, XXX−XXXG

diverted rats). The results suggest that cinnarizine absorptionfrom lipid colloidal phases is relatively insensitive to thephysical nature of the colloidal milieu, and is instead controlledlargely by Cfree.

Bile-Mediated Dilution of Cinnarizine-Loaded Mi-celles and Vesicles Generates Drug Supersaturation.The equilibrium solubility and change in cinnarizine solubiliza-tion capacity of model micelles and vesicles before and after 1:1v/v dilution with bile, bile pH 6.30 or buffer pH 6.30 aretabulated in Table 4, and shown graphically in Figure 5.Dilution of the micellar or vesicular systems in a 1:1 v/v ratio

with bile obtained from donor animals led to significantdecreases in cinnarizine solubilization capacity. As a proportionof initial, the solubilization capacity of the micellar systemdropped significantly from 100% to 14%, and for vesicles thedecrease was even greater from 100% to 7%. This was in spiteof the fact that the bile salt concentration in donor bile washigher than that in the micellar or vesicular system (averagetotal bile salt concentration in donor bile was 14.7 ± 0.9 mM;mean ± SEM, n = 13), and therefore “dilution” with bile didnot reduce bile salt concentrations below the critical micellarconcentration (CMC) and instead increased the overall bile saltconcentration.The pH of donor bile was higher than intestinal pH (average

pH of donor bile was 8.02 ± 0.02; mean ± SEM, n = 5). Assuch additional studies were performed to examine whether theeffects on solubility reflected a pH effect. 1:1 v/v dilution of thephases with pH adjusted bile (pH of bile adjusted to 6.30 tomatch the pH of the micelles and vesicles) also resulted insignificant decreases in solubilization capacity (to 28% and 15%

Table 3. Cinnarizine Disappearance Papp (×106 cm/s) fromthe Intestinal Perfusate, Appearance Papp (×106 cm/s) in theMesenteric Blood and Steady-State Absorptive Flux intoMesenteric Blood (ng/5 min/10 cm2) after 60 min of Single-Pass Perfusion of ∼10 cm2 Segments of Rat Jejunum withModel Micelles and Vesicles, with and without 1:1 v/vCoperfusion with Rat Bile, Rat Bile pH 6.30 or Buffer pH6.30a

SSratiob

disappearancePapp (×10

6

cm/s)

appearancePapp (×10

6

cm/s)

flux into mesentericblood (ng/5 min/10

cm2)

micelles 0.4 19.4 ± 3.0 1.1 ± 0.2 310 ± 50micelles +bileSSc

6.1 29.3 ± 3.0d 3.7 ± 0.1d 979 ± 30d

micelles +bile pH6.30SS

3.0 39.5 ± 5.6d 4.2 ± 1.0d 1099 ± 279d

micelles +buffer pH6.30SS

2.1 38.9 ± 4.1d 4.4 ± 0.4d 1180 ± 100d

vesicles 0.4 22.0 ± 2.2 1.2 ± 0.1 340 ± 29vesicles +bileSS

11.6 66.0 ± 18.3e 2.0 ± 0.7 499 ± 155

aValues calculated using data obtained after steady-state attainment (t= 40−60 min). In all experiments, cinnarizine concentration inperfusate and total perfusate flow rate were kept constant at 0.1 mg/mL and 0.1 mL/min, respectively. Data represent mean ± SEM of n =3−4 experiments. bSS ratio = supersaturation ratio = (supersaturated)concentration of drug in perfusate/equilibrium solubility of drug inperfusate. cSS denotes drug supersaturation in perfusate. dSignificantincrease from micelles alone. eSignificant increase from vesicles alone.

Figure 3. (A) Absorptive flux of cinnarizine (CIN) into mesenteric blood (ng/5 min/10 cm2) and (B, C) CIN disappearance from intestinalperfusate (% drug dose passing through jejunum) when micelles/vesicles were perfused through an isolated rat jejunal segment (∼10 cm2), with(open symbols) and without (filled symbols) 1:1 v/v coperfusion with rat bile. Coperfusion of micelles and vesicles with donor bile generates drugsupersaturation in situ within the perfused jejunal segment. SS denotes experiments where drug is supersaturated in the perfusate. The degree of drugprecipitation within the perfusate is illustrated in B and C as the difference in perfusate concentration between pre- and postcentrifugation data.Precipitation was significant in the vesicle, but not micellar groups. Experiments were performed using an in situ single-pass rat jejunum perfusionmodel. In all experiments, the concentration of cinnarizine in perfusate and the total perfusate flow rate were kept constant at 0.1 mg/mL and 0.1mL/min, respectively. Data represent mean ± SEM of n = 3−4 experiments.


dx.doi.org/10.1021/mp3006566 | Mol. Pharmaceutics XXXX, XXX, XXX−XXXH

of initial for micelles and vesicles, respectively), although thedecrease was slightly attenuated when compared to non-pHadjusted bile. Finally, the micellar and vesicular systems werediluted 1:1 v/v with buffer at pH 6.30 in an attempt touncouple simple dilution effects from pH effects and bile-mediated effects. Dilution with pH 6.30 buffer also decreasedcinnarizine solubilization capacity significantly, although to alesser extent (41% and 21% in micelles and vesicles,respectively).Analysis of the kinetics of drug precipitation (Figure 6)

demonstrated that bile addition to cinnarizine-loaded micelles

Figure 4. Systemic plasma concentration−time profiles of cinnarizine (CIN) following a 2 h intraduodenal infusion of cinnarizine-loaded (0.2 mg/mL) micelles (left panel) and vesicles (right panel) to bile-intact and bile-diverted rats. Consistent with observations in rat jejunum perfusion studies,bile-induced supersaturation translated into increased in vivo exposure during the absorption phase in the case of micelles but not vesicles. Datarepresent mean ± SEM of n = 4 rats. Statistical significance (p < 0.05) is denoted by the symbol *.

Table 4. Equilibrium Solubility (37 °C) and Percent OriginalSolubilization Capacity Values of Cinnarizine in ModelMicelles and Vesicles, before and after 1:1 v/v Addition ofRat Bile, Rat Bile pH 6.30 or Buffer pH 6.30a

equilibrium solubility(μg/mL)

percent original solubilizationcapacity (%)b

Micellesmicelles alone 236 ± 9.3 100

+ bile (1:1) 16.5 ± 0.8 14.0 ± 0.7+ bile pH6.30 (1:1)

33.1 ± 1.0 28.0 ± 0.9

+ buffer pH6.30 (1:1)

48.7 ± 0.8 41.2 ± 0.6

Vesiclesvesicles alone 232 ± 3.3 100

+ bile (1:1) 8.60 ± 0.5 7.4 ± 0.4+ bile pH6.30 (1:1)

17.4 ± 0.4 15.0 ± 0.3

+ buffer pH6.30 (1:1)

24.9 ± 0.3 21.4 ± 0.2

aData represent mean ± SEM of n = 3−4 determinations. bPercentoriginal solubilization capacity = (solubilityfinal × volumefinal/solubilityinitial × volumeinitial) × 100%.

Table 5. Pharmacokinetic Parameters for Cinnarizine (CIN) after Intraduodenal Administration of Cinnarizine-Loaded (0.2mg/mL) Micelles and Vesicles to Bile-Intact and Bile-Diverted Ratsa

experiment group formulation type CIN dose (mg/kg) AUC0−8h (ng h/mL) Cmax (ng/mL) Tmax (h)

micellesbile-intact micelles 2 502 ± 59b 138.8 ± 11.6b 1.6 ± 0.2bile-diverted micelles 2 362 ± 39 85.7 ± 14.3 2.0 ± 0.4

vesiclesbile-intact vesicles 2 411 ± 74 109.4 ± 10.6 2.3 ± 0.3bile-diverted vesicles 2 393 ± 82 97.6 ± 12.8 2.0 ± 0.4

aValues represent mean ± SEM of n = 4 rats. bSignificant increase when compared to micelles (bile-diverted) group.

Figure 5. Percent original cinnarizine (CIN) solubilization capacity ofmodel colloids (micelles or vesicles), before and after a 1:1 v/vaddition of rat bile, rat bile pH 6.30 or buffer pH 6.30. Data representmean ± SEM of n = 3−4 determinations. Statistical significantdifference (p < 0.05) to colloid only is denoted by the symbol *.

Figure 6. Kinetics of cinnarizine (CIN) precipitation from modelmicelles (filled circles, n = 5) and model vesicles (open circles, n = 4)upon addition of rat bile (in a 1:1 v/v ratio) at t = 0. Addition of bilereduces the equilibrium cinnarizine solubilization capacity of micellesand vesicles to 14% and 7% of initial, respectively (see Table 4, andshown here as the lines denoted equilibrium solubility). Cinnarizinesupersaturation appeared to be maintained for longer in micelles thanin vesicles. Cinnarizine was loaded into micelles and vesicles at 80%saturation (∼0.2 mg/mL). Each line represents individual experiments.


dx.doi.org/10.1021/mp3006566 | Mol. Pharmaceutics XXXX, XXX, XXX−XXXI

and vesicles (cinnarizine present in phases at 0.2 mg/mL,∼80% saturated solubility) did not result in immediate drugprecipitation, and was preceded by a period of supersaturation.The time taken for cinnarizine precipitation to occur wasvariable, however supersaturation was maintained for longerperiods in micelles (>20 min in all cases) when compared tovesicles (1−20 min).Bile-Induced Drug Supersaturation Increases Jejunal

Absorptive Flux for Micelles but Not Vesicles. Theabsorptive flux vs time profiles for cinnarizine under super-saturated conditions (i.e., when bile was coperfused with thephases) and under nonsupersaturated conditions (when bilewas not coperfused with the phases) are shown in Figure 3A.Steady-state absorptive flux, disappearance Papp and appearancePapp of cinnarizine in all perfusion experiments are reported inTable 3.Coperfusion of micelles with bile in a 1:1 v/v ratio increased

the absorptive flux, disappearance Papp and appearance Papp ofcinnarizine from micelles 3.2-fold, 1.5-fold and 3.4-fold,respectively. In contrast, 1:1 v/v coperfusion of vesicles withbile did not lead to significant changes in cinnarizine absorptiveflux or appearance Papp. Disappearance Papp for cinnarizine didincrease 3.0-fold when vesicles were coperfused with bile,however the drop in perfusate drug concentration was largely aresult of rapid drug precipitation in the perfusate (see below).For the micellar preparation, bile-induced supersaturation

was relatively stable throughout the experimental period. Thusthe cinnarizine concentration in the perfusate was essentiallythe same before or after centrifugation (Figure 3B). In contrast,when vesicles were coperfused with bile, significant drugprecipitation was observed during the time required forperfusate to transit the jejunal segment (as indicated by thedifference between the pre- and postcentrifugation data inFigure 3C). This was consistent with the in vitro dilutionprofiles in Figure 6.To distinguish between bile-induced increases in cinnarizine

absorptive flux resulting from drug supersaturation, pHincreases (since an increase in pH might be expected toincrease the permeability of a weak base) and nonspecificeffects of bile components on membrane permeability, micelleswere also coperfused with pH-adjusted bile (pH 6.30) andbuffer (pH 6.30). 1:1 v/v coperfusion of micelles with bile pH6.30 or buffer pH 6.30 increased cinnarizine absorptive flux 3.5-fold and 3.8-fold, respectively, when compared to the perfusionof micelles alone (Figure 7, Table 3). Stable supersaturationwas also generated within the perfused jejunal segment in theseexperiments (Figure S2 in the Supporting Information:cinnarizine concentration in the perfusate was essentially thesame before and after centrifugation).Since 1:1 v/v coperfusion of micelles with bile, bile pH 6.30

and buffer pH 6.30 are all expected to generate cinnarizinesupersaturation within the perfused jejunal segment (accordingto the solubility data in Table 4, and the lack of drugprecipitation in outflow perfusate in all cases), the observationthat absorptive flux enhancement was similar in all groups(Figure 7) suggests that the enhancement was attributable todrug supersaturation, and not an increase in system pH (bycomparing micelles + bile group with micelles + bile pH 6.30group), or nonspecific effects of bile on membrane permeability(by comparing micelles + bile pH 6.30 group with micelles +buffer pH 6.30 group). However, the degree of fluxenhancement did not correlate with the degree of super-saturation, as flux enhancement was similar in all groups despite

the significantly higher supersaturation ratio generated in themicelles + bile group, when compared to the micelles + bile pH6.30 group and micelles + buffer pH 6.30 group (super-saturation ratio of 6 vs 2−3) (Table 3). Previous studies haveshown that, for lipophilic drugs such as cinnarizine, the drugfraction extracted into octanol (and analogous to the permeablefraction) as a function of pH is shifted to lower pHs than wouldbe expected based on the un-ionized fraction.41,42 This left shiftwould limit pH effects on permeability over the range of pH 6−8, consistent with the observations here.

Bile-Induced Drug Supersaturation Increases in VivoCinnarizine Exposure after Intraduodenal Infusion. Thesystemic plasma concentration−time profiles of cinnarizinefollowing intraduodenal infusion of cinnarizine-loaded micellesand vesicles (0.2 mg/mL; ∼80% saturated solubility) to bile-intact and bile-diverted rats are shown in Figure 4. Thepharmacokinetic parameters of cinnarizine for the bioavail-ability studies are reported in Table 5. Consistent with theresults from the rat jejunal permeability studies, intraduodenaladministration of micelles to bile-intact rats resulted insignificantly higher systemic plasma concentrations of cinnar-izine at t = 1, 1.5 and 2 h (AUC0−8h increased 1.4-fold) whencompared to bile-diverted rats. In contrast, when vesicles weredosed to bile-intact rats, the systemic plasma concentration−time profiles and AUC0−8h of cinnarizine were not differentfrom that observed in bile-diverted rats.

■ DISCUSSIONAfter oral administration, the absorption of poorly water-soluble drugs (PWSD) is often limited by slow dissolution andlow solubility in the GI tract. LBF overcome many of thedissolution limitations of PWSD (by providing a mechanism tocircumvent dissolution from the solid to liquid state), however,the solubility limitations of PWSD are seemingly unaddressed,

Figure 7. Absorptive flux−time profiles of cinnarizine (CIN) whenmicelles were perfused through an isolated rat jejunal segment (∼10cm2), with and without 1:1 v/v coperfusion with rat bile, rat bile pH6.30 or buffer pH 6.30. SS denotes experiments where drug issupersaturated in perfusate. Coperfusion of rat bile, rat bile pH 6.30 orbuffer pH 6.30 with micelles generates drug supersaturation in situwithin the perfused jejunal segment in all cases, and increasedcinnarizine absorptive flux by 3.2-fold, 3.5-fold and 3.8-fold,respectively. Experiments were performed using an in situ single-passrat jejunum perfusion model. In all experiments, the concentration ofcinnarizine in perfusate and the total perfusate flow rate were keptconstant at 0.1 mg/mL and 0.1 mL/min, respectively. This series ofexperiments provides further support for the suggestion thatsupersaturation was responsible for the increase in absorptive fluxseen in Figure 3A. Data represent mean ± SEM of n = 3−4experiments.


dx.doi.org/10.1021/mp3006566 | Mol. Pharmaceutics XXXX, XXX, XXX−XXXJ

since solubilization in the lipid-based colloidal phases that resultfrom the digestion of LBF does not typically enhance free drugconcentrations.2,43 Nonetheless, coadministration of lipids(either formulation lipids or via coadministration with lipid-rich foods) remains a highly effective means to promote theabsorption of PWSD. This suggests the potential for alternativemechanisms by which LBF enhance drug absorption. In thecurrent communication, two possible alternatives to thetraditional model of drug absorption from LBF have beenexplored: first, that drug absorption from lipid colloidal phasesmay involve a collisional uptake component (i.e., drugabsorption directly from the solubilized phase); and second,that flux across the absorptive membrane may be enhanced by atransient increase in the thermodynamic activity of drug inintestinal colloidal phases due to supersaturation.The data describing drug absorption from micellar and

vesicular colloidal phases suggest that direct interactionsbetween colloids (or at least the systems examined here) andthe absorptive membrane do not play an important role incinnarizine absorption. Thus, comparable absorptive flux(Figure 3A: filled symbols) and systemic plasma concen-tration−time profiles (Figure 4: bile-diverted rats) wereobserved following jejunal perfusion and intraduodenal infusionof micelles and vesicles. The thermodynamic activity and Cfreeof both systems were held constant, but the large difference inhydrodynamic radius (9 nm of micelles vs 443 nm of vesicles)and higher bile salt, LPC and lipid concentrations in themicellar system (Table 1) suggest that the number ofadministered micellar particles was substantially higher thanthat of vesicular particles. Collision-mediated absorption ishighly sensitive to increased particle number, since thisincreases the statistical likelihood of collisions and collisionaltransfer.7 As such collisional interactions did not appear todictate the degree of drug absorption from micelles and vesiclesand, instead, absorption was seemingly controlled bythermodynamic activity (or Cfree). Differences in particle sizemight also be expected to alter colloid diffusion across theunstirred water layer (UWL). The similarity in absorptionprofiles from micelles and vesicles (Figure 3A: filled symbols)therefore suggests that diffusion across the UWL either wasnonlimiting or was not affected by particle size in thisexperimental model. Similarly, the rates of replenishment ofCfree (i.e., the rate of re-establishment of the equilibriumbetween solubilized and free drug) might be expected to bedifferent between the two different particle size colloids,however this was presumably sufficiently fast in both cases tohave little impact on drug absorption in the current model.Data obtained from the lipid uptake receptor inhibition

studies further support the notion that collisional uptakemechanisms have a limited role in drug absorption from thesesystems. SR-BI, CD36 and NPC1L1 were examined since theyhave previously been shown to mediate the cellular uptake offatty acids and/or cholesterol,12,15,44 the absorption of which isalso facilitated by micellar solubilization. Whether SR-BI, CD36and NPC1L1 function as authentic transporters that directlymediate lipid absorption,12−15,44 or whether they act tofacilitate intracellular lipid trafficking or to modify signalingprocesses that mediate lipid absorption,45−47 or both, remainscontentious, but nonetheless all merit examination here. Recentreports also suggest the involvement of SR-BI, CD36 andNPC1L1 in the absorption of fat-soluble nutrients (such ascarotenoids,15,48 vitamin D49 and vitamin E50), providingfurther support for a role in drug absorption. In the current

study, consistent with previous reports in vivo, inhibition ofNPC1L1 reduced cholesterol absorption (Figure 2B),14 andinhibition of CD36 reduced (albeit nonsignificantly) theabsorption of oleic acid (Figure 2C).12 In contrast, inhibitionof SR-BI and CD36 had little impact on cholesterol absorption(Figure 2B). Inhibition of SR-BI and CD36 was expected toreduce cholesterol absorption based on previous in vitrostudies.12,13,15 However, in vivo evidence of a role of SR-BIand CD36 in cholesterol absorption is less clear and, forexample, no significant differences in intestinal cholesterolabsorption were reported in SR-BI knockout vs wild-typemice.15,51 The lack of effect of SR-BI and CD36 inhibition oncholesterol absorption in the current study therefore runscontrary to previous in vitro studies but is in agreement withsome previous in vivo data. Inhibition of endocytosis pathwaysvia administration of monensin also failed to reduce cholesteroland cinnarizine absorption. Rather, the volume of distributionand clearance of cholesterol and cinnarizine appeared to bereduced, resulting in increases in plasma exposure. Sinceexogenously dosed 14C-cholesterol and cinnarizine are likely tobe present within lipoproteins in the systemic circulation, thechanges in systemic disposition of 14C-cholesterol andcinnarizine in the monensin-treated rats may reflect inhibitionof receptor-mediated endocytosis of LDL.52

Inhibition of SR-BI, CD36 and NPC1L1 did not result insignificant changes to the systemic exposure of cinnarizinefollowing intraduodenal infusion of a lipid emulsion for-mulation in rats (Figure 2A). The data suggest that while lipidcolloidal phases may be capable of direct interaction with lipidreceptors (as suggested previously), the transfer of solubilizedcontent into absorptive cells is likely selective and moreapplicable to solubilized lipids and nutrients rather than drugs.Together with the data describing cinnarizine absorption frommicelles and vesicles, the results indicate that cinnarizineabsorption from lipid colloidal phases is largely independent ofthe physical nature of the infused colloid (realizing that in thesefirst experiments cinnarizine was present at concentration wellbelow the solubilization limit and therefore under conditionswhere solubility/precipitation-mediated events were avoided),is not influenced significantly by common lipid uptakereceptors and, instead, appears to be primarily dependent onthe free drug concentration in equilibrium with the solubilizedreservoir.Subsequent studies addressed the possibility that bile

secretion may enhance drug absorption via the induction ofsupersaturation during the intestinal processing of dietary orformulation-related lipids. The addition of bile to cinnarizine-loaded micelles led to sustained drug supersaturation thatultimately led to increased intestinal drug absorption andsystemic drug exposure. In contrast, although addition of bile tocinnarizine-loaded vesicles also led to supersaturation, themetastable supersaturated state was less stable than thatgenerated by bile addition to micelles, resulting in more rapidprecipitation of solubilized drug and therefore a lack of increasein drug absorption and systemic drug exposure. The dataindicate that bile-mediated dilution of lipid colloidal phases mayrepresent an endogenous mechanism of supersaturationgeneration during lipid processing in the small intestine; andthat the transient increase in thermodynamic activity may leadto enhanced drug absorption. The observations also highlightboth the potential for supersaturation to enhance drugabsorption and the need to achieve an optimal balance betweendrug supersaturation and drug precipitation.


dx.doi.org/10.1021/mp3006566 | Mol. Pharmaceutics XXXX, XXX, XXX−XXXK

Interaction of intestinal colloidal phases with bile leads todilution, an increase in pH and an increase in theconcentrations of bile components (bile salts, phosphatidylcho-line and cholesterol) associated with the colloidal species. For asolubilized system above the critical micellar concentration(CMC), simple 1:1 v/v dilution is expected to reduce thesolubilized concentration, but to maintain total solubilizationcapacity (i.e., for the drug concentration to drop by 50% butthe volume to double and therefore for solubilization capacityto remain unchanged). However, the data in Table 4 show thatdilution of the micelles or vesicles with bile or buffer results in adrop in total solubilization capacity to only 7−41% of initial.Greater proportional decreases were apparent for dilution ofvesicles (in all cases when compared to micelles), and afterdilution with bile rather than buffer (Table 4). The loss ofsolubilization capacity on dilution suggests the likelihood of aphase transition to structures with reduced solubilizationcapacity. Although a complete explanation for these phasetransitions is not apparent at this time, for the micellar systems,it may be related to the ability of ionized caprylic acid to self-associate and form fatty acid micelles at high lipidconcentration.53 Thus, dilution of medium-chain colloidalphases may reduce the concentration of caprylic acid belowthe CMC, leading to a loss of cinnarizine solubilization capacityin the micellar phase. For vesicles, previous studies havesuggested that increasing bile concentrations facilitate avesicular to micellar transition.54,55 Since micelles are expectedto have lower solubilization capacities for lipophilic drugs thanvesicles (micelles are smaller and less lipid-rich),22,56 areduction in cinnarizine solubilization capacity might thereforebe anticipated when bile is added to model vesicles. In the caseof the micelles, therefore, it seems likely that the addition ofbile disrupted the structure of swollen, mixed micelles leadingto lower colloidal lipid content and lower solubilizationcapacity. Unfortunately, attempts to quantify changes toparticle size on bile addition were unsuccessful due to highpolydispersity. However, the broad trends observed wereconsistent with the suggestions above and decreases in particlesize were apparent for vesicles (consistent with initiation of a

vesicle to micelle transition) and increases in particle size forthe mixed micelles (consistent with micellar destabilization andtransformation to less dispersed structures) (data not shown).An additional complexity in these dilution studies was therealization that the pH of bile (pH 8.02) is higher than thatnormally employed for simulated intestinal fluids,23,57 andtherefore incubation of the colloidal phases (prepared at pH6.30) with bile increases system pH. Cinnarizine is a weak basewith a pKa of 7.47

58 and is therefore expected to be less ionized,less soluble and potentially more permeable at higher pHs. Theimpact of pH in the current studies was therefore examined bydilution of micelles and vesicles with pH-corrected bile at pH6.30. Comparison to the data obtained with bile at pH 8.02suggests that the higher pH of bile provides an additional driverfor drug precipitation/supersaturation since the solubility dropwas greater after incubation with bile at pH 8.02 vs bile at pH6.30While dilution of micelles and vesicles with bile resulted in

decreases in drug solubilization capacity, cinnarizine did notprecipitate immediately. A period of drug supersaturation wasevident for both micellar and vesicular systems, although drugprecipitation from vesicles was much more rapid than frommicelles. This in turn translated into increases in drugabsorption (Figure 3A) and systemic exposure (Figure 4A)for the micellar systems. The difference in the capacity ofmicelles and vesicles to maintain drug supersaturation (Figure6) may be explained by the difference in the degree of drugsupersaturation induced by dilution with bile. The degree ofsupersaturation is described by the supersaturation ratio, whichis the ratio of the (supersaturated) concentration of drug insolution relative to the equilibrium solubility of the drug in thesame system.16 Crystallization theory suggests that thethermodynamic drivers of precipitation from supersaturatedsolutions increase with increasing supersaturation ratios, as thelikelihood of nucleation and crystal growth increases withincreases in (metastable) drug concentration in solution.16,59

Here, 1:1 v/v addition of bile resulted in the attainment ofcinnarizine supersaturation ratios of 6 and 12, in micelles andvesicles, respectively. Therefore, the faster rate of drug

Figure 8. The dual role of bile during lipid digestion and dispersion. (i) Bile-mediated solubilization of lipid digestion products at the interface of adigesting oil droplet results in the generation of lipid colloidal phases such as vesicles and micelles that maintain drug solubilization in the smallintestine. (ii) The continuing interaction of secreted bile with existing lipid colloidal phases in the lumen results in progressively less lipid-rich phaseswith lowered solubilization capacity. Thus, ongoing bile-mediated dilution of lipid colloidal phases promotes drug supersaturation and enhances drugabsorption by increasing drug thermodynamic activity in colloidal phases. The combination of the two highly kinetic, bile-mediated events affords ameans to simultaneously increase solubilization capacity and promote thermodynamic activity of coadministered drug in the small intestine, and maycontribute to the increase in drug absorption often observed with lipid coadministration. D represents the free concentration of drug available forabsorption. Dss is used to signify the increase in free concentration resulting from bile-mediated supersaturation that drives increases in drugabsorption.


dx.doi.org/10.1021/mp3006566 | Mol. Pharmaceutics XXXX, XXX, XXX−XXXL

precipitation from vesicles (when compared to micelles) maybe explained by the higher supersaturation ratio induced by bileaddition. It is also possible that micellar structures are moreeffective in stabilizing supersaturation when compared tovesicular structures, although this has not been examinedexplicitly here. Notably, cinnarizine was found to precipitate inthe crystalline form in these experiments (Figure S3 in theSupporting Information), precluding the possibility that theenhanced cinnarizine absorption from micelles (when bile wascoperfused) observed in Figure 3A was due to acceleratedcinnarizine dissolution from precipitated amorphous forms.The role of bile in enhancing drug absorption from LBF has

been reported previously.60−62 In the majority of cases, bile-mediated bioavailability enhancement has been suggested tostem from the ability of bile to expand the solubilizationreservoir for PWSD in the GI tract. This is typically assumed tooccur via PWSD solubilization in simple bile micelles, or via theability of bile to solubilize lipid digestion products and togenerate more complex lipid colloidal phases with enhancedsolubilization capacities. It seems likely that the ability tosolubilize lipid digestion products and to promote colloidformation remains an integral part of the role of bile insupporting drug (and lipid) absorption. However, the datadescribed here suggest that continued dilution of lipid colloidalphases with bile in the small intestinal lumen may also lead tophysical changes that promote drug supersaturation, andultimately promote drug absorption. In doing so, super-saturation induction may be a means by which the decreasein thermodynamic activity inherent in solubilization is reversed,such that the free concentration of drug available for absorptionis maximized. Thus, a dual role of bile in facilitating drugabsorption from LBF may be conceived (see Figure 8). First,bile-mediated solubilization of lipid digestion products at theinterface of a digesting lipid droplet results in the generation oflipid colloidal phases such as vesicles and micelles that promotedrug solubilization during lipid digestion. Second, continuedbile-mediated dilution of existing lipid colloidal phasespromotes drug supersaturation, and enhances drug absorptionby significantly increasing drug thermodynamic activity incolloidal phases. The combination of these two highly kineticevents likely contributes to the drug absorption benefits oftenobserved with lipid coadministration, as it affords a means tosimultaneously increase solubilization capacity and promotethermodynamic activity of coadministered PWSD in the smallintestine.Supersaturation induction via interaction with bile also

provides a means of overcoming the recently describedsolubility−permeability interplay observed in studies wherePWSD are coadministered with cosolvents, cyclodextrins orsurfactant systems.5,6,63 In these studies the authors describethe reduction in thermodynamic activity common to mostsolubilization technologies and show that this offsets thepotential increases in membrane flux that might be expected byan increase in solubilized drug concentration.5,6 More recentstudies by the same authors have shown that this solubility−permeability interplay can be addressed via the use ofamorphous solid dispersion formulations that stimulate super-saturation, but do not promote solubilization.64,65 Here wereport that essentially similar outcomes are also possible withsolubilizing formulations, when the solubilizing formulationscontain lipids and when the kinetic changes that occur in the GIlumen in the presence of bile secretion promote super-saturation. The current data therefore suggest that endogenous

lipid processing pathways provide an exquisitely sensitive andtriggerable supersaturation mechanism that allows drug toremain in a solubilized state during initial lipid digestion and athigh lipid:bile concentration ratios, but that ongoing bilesecretion subsequently provides a boost to thermodynamicactivity and in doing so supports enhanced absorption. This isin contrast to other common solubilization strategies that maynot interact with the dynamic GI environment, or for whichinteraction with bile typically reduces thermodynamic activityby increasing solubilization capacity. Conversely, the ability ofLBF to promote drug solubilization until drug supersaturationis triggered in the small intestine may confer an advantage overformulation approaches that only utilize supersaturatingstrategies, as the risk of drug precipitation in the GI tractmay be reduced by solubilization within lipid colloidal phases.

■ CONCLUSION

Improved understanding of the mechanism of drug absorptionfrom lipid colloidal phases such as micelles and vesicles isrequired to provide a platform for more rational design of lipid-based formulations. Using medium-chain lipids, we havedemonstrated that the absorption of cinnarizine (a lipophilic,poorly water-soluble drug) may be enhanced when drugsupersaturation is generated during bile-mediated dilution oflipid colloidal phases. Previous studies suggest that a similarinduction of supersaturation may occur as a result of initiationof digestion of some microemulsion-based LBFs.21 Theseobservations indicate that supersaturation and its associatedbenefits in enhancing drug absorption may occur intrinsicallyduring LBF incorporation into endogenous lipid processingpathways in the small intestine. Future work will be directedtoward assessing the impact of bile dilution on supersaturationtendency, and thus absorption, for an extended range of PWSDand in a series of different micellar and vesicular colloidalsystems.

■ ASSOCIATED CONTENT

*S Supporting InformationTable S1 showing pharmacokinetic parameters for cinnarizineafter intravenous administration in rats. Figure S1 showingsystemic plasma concentration−time profiles of cinnarizine and14C-labeled cholesterol following intravenous infusion in rats.Figure S2 showing perfusate disappearance profiles ofcinnarizine when micelles were perfused through an isolatedrat jejunal segment, with and without 1:1 v/v coperfusion withrat bile pH 6.30 or buffer pH 6.30. Figure S3 showing polarizedlight microscopy images of the crystalline cinnarizineprecipitate following precipitation kinetics experiments. Thismaterial is available free of charge via the Internet at http://pubs.acs.org.

■ AUTHOR INFORMATION

Corresponding Author*N.L.T.: phone, +61 3 9903 9138; fax, +61 3 9903 9583; e-mail, [email protected]. C.J.H.P.: phone, +61 39903 9649; fax, +61 3 9903 9583; e-mail, [email protected].

NotesThe authors declare no competing financial interest.


dx.doi.org/10.1021/mp3006566 | Mol. Pharmaceutics XXXX, XXX, XXX−XXXM

http://pubs.acs.org

http://pubs.acs.org

mailto:[email protected]



■ ACKNOWLEDGMENTS

Funding support from the National Health and MedicalResearch Council (NHMRC) is gratefully acknowledged.

■ ABBREVIATIONS USED

LBF, lipid-based formulations; PWSD, poorly water-solubledrugs; GI, gastrointestinal; LPC, L-α-lysophosphatidylcholine;CD36, cluster of differentiation 36; SR-BI, scavenger receptorclass B type 1; NPC1L1, Niemann-Pick C1 like 1; BLT-1, blocklipid transport-1; SSO, sulfo-N-succinimidyl oleate; iv, intra-venous; HPLC, high performance liquid chromatography;NMR, nuclear magnetic resonance; DMSO, dimethyl sulfoxide;TBME, tert-butyl methyl ether; SEIF, simulated endogenousintestinal fluid; SEM, standard error of the mean; CMC, criticalmicellar concentration; UWL, unstirred water layer; HDL,high-density lipoprotein; LDL, low-density lipoprotein; CIN,cinnarizine; Ch, cholesterol; OA, oleic acid; SS ratio, super-saturation ratio

■ REFERENCES(1) Williams, H. D.; Trevaskis, N. L.; Charman, S. A.; Shanker, R. M.;Charman, W. N.; Pouton, C. W.; Porter, C. J. H. Strategies to addresslow drug solubility in discovery and development. Pharmacol. Rev.2013, 65 (1), 315−499.(2) Poelma, F. G. J.; Breas, R.; Tukker, J. J.; Crommelin, D. J. A.Intestinal absorption of drugs. The influence of mixed micelles on thedisappearance kinetics of drugs from the small intestine of the rat. J.Pharm. Pharmacol. 1991, 43 (5), 317−324.(3) Katneni, K.; Charman, S. A.; Porter, C. J. H. Permeabilityassessment of poorly water-soluble compounds under solubilizingconditions: The reciprocal permeability approach. J. Pharm. Sci. 2006,95 (10), 2170−2185.(4) Amidon, G. E.; Higuchi, W. I.; Ho, N. F. H. Theoretical andexperimental studies of transport of micelle-solubilized solutes. J.Pharm. Sci. 1982, 71 (1), 77−84.(5) Miller, J. M.; Beig, A.; Krieg, B. J.; Carr, R. A.; Borchardt, T. B.;Amidon, G. E.; Amidon, G. L.; Dahan, A. The solubility−permeabilityinterplay: Mechanistic modeling and predictive application of theimpact of micellar solubilization on intestinal permeation. Mol.Pharmaceutics 2011, 8 (5), 1848−1856.(6) Dahan, A.; Miller, J. M.; Hoffman, A.; Amidon, G. E.; Amidon, G.L. The solubility−permeability interplay in using cyclodextrins aspharmaceutical solubilizers: Mechanistic modeling and application toprogesterone. J. Pharm. Sci. 2010, 99 (6), 2739−2749.(7) Wootan, M. G.; Bernlohr, D. A.; Storch, J. Mechanism offluorescent fatty acid transfer from adipocyte fatty acid binding proteinto membranes. Biochemistry 1993, 32 (33), 8622−8627.(8) Thumser, A.; Tsai, J.; Storch, J. Collision-mediated transfer oflong-chain fatty acids by neural tissue fatty acid-binding proteins(FABP). J. Mol. Neurosci. 2001, 16 (2), 143−150.(9) Yano, K.; Masaoka, Y.; Kataoka, M.; Sakuma, S.; Yamashita, S.Mechanisms of membrane transport of poorly soluble drugs: Role ofmicelles in oral absorption processes. J. Pharm. Sci. 2010, 99 (3),1336−1345.(10) Gao, P.; Morozowich, W. Development of supersaturatable self-emulsifying drug delivery system formulations for improving the oralabsorption of poorly soluble drugs. Expert Opin. Drug Delivery 2006, 3(1), 97−110.(11) Gu, X.; Trigatti, B.; Xu, S.; Acton, S.; Babitt, J.; Krieger, M. Theefficient cellular uptake of high density lipoprotein lipids via ScavengerReceptor Class B Type I requires not only receptor-mediated surfacebinding but also receptor-specific lipid transfer mediated by itsextracellular domain. J. Biol. Chem. 1998, 273 (41), 26338−26348.(12) Nassir, F.; Wilson, B.; Han, X.; Gross, R. W.; Abumrad, N. A.CD36 is important for fatty acid and cholesterol uptake by the

proximal but not distal intestine. J. Biol. Chem. 2007, 282 (27),19493−19501.(13) Hauser, H.; Dyer, J. H.; Nandy, A.; Vega, M. A.; Werder, M.;Bieliauskaite, E.; Weber, F. E.; Compassi, S.; Gemperli, A.; Boffelli, D.;Wehrli, E.; Schulthess, G.; Phillips, M. C. Identification of a receptormediating absorption of dietary cholesterol in the intestine.Biochemistry 1998, 37 (51), 17843−17850.(14) Altmann, S. W.; Davis, H. R.; Zhu, L.; Yao, X.; Hoos, L. M.;Tetzloff, G.; Iyer, S. P. N.; Maguire, M.; Golovko, A.; Zeng, M.; Wang,L.; Murgolo, N.; Graziano, M. P. Niemann-Pick C1 Like 1 protein iscritical for intestinal cholesterol absorption. Science 2004, 303 (5661),1201−1204.(15) Van Bennekum, A.; Werder, M.; Thuahnai, S. T.; Han, C.-H.;Duong, P.; Williams, D. L.; Wettstein, P.; Schulthess, G.; Phillips, M.C.; Hauser, H. Class B Scavenger Receptor-mediated intestinalabsorption of dietary β-carotene and cholesterol. Biochemistry 2005,44 (11), 4517−4525.(16) Brouwers, J.; Brewster, M. E.; Augustijns, P. Supersaturatingdrug delivery systems: The answer to solubility-limited oralbioavailability? J. Pharm. Sci. 2009, 98 (8), 2549−2572.(17) Porter, C. J. H.; Kaukonen, A. M.; Boyd, B. J.; Edwards, G. A.;Charman, W. N. Susceptibility to lipase-mediated digestion reducesthe oral bioavailability of danazol after administration as a medium-chain lipid-based microemulsion formulation. Pharm. Res. 2004, 21(8), 1405−1412.(18) Kaukonen, A. M.; Boyd, B.; Porter, C.; Charman, W. Drugsolubilization behavior during in vitro digestion of simple triglyceridelipid solution formulations. Pharm. Res. 2004, 21 (2), 245−253.(19) Anby, M. U.; Williams, H. D.; McIntosh, M.; Benameur, H.;Edwards, G. A.; Pouton, C. W.; Porter, C. J. H. Lipid digestion as atrigger for supersaturation: Evaluation of the impact of supersaturationstabilization on the in vitro and in vivo performance of self-emulsifyingdrug delivery systems. Mol. Pharmaceutics 2012, 9 (7), 2063−2079.(20) Cuine, J.; Charman, W.; Pouton, C.; Edwards, G.; Porter, C.Increasing the proportional content of surfactant (Cremophor EL)relative to lipid in self-emulsifying lipid-based formulations of danazolreduces oral bioavailability in beagle dogs. Pharm. Res. 2007, 24 (4),748−757.(21) Pouton, C. W.; Porter, C. J. H. Formulation of lipid-baseddelivery systems for oral administration: Materials, methods andstrategies. Adv. Drug Delivery Rev. 2008, 60 (6), 625−637.(22) Kossena, G. A.; Boyd, B. J.; Porter, C. J. H.; Charman, W. N.Separation and characterization of the colloidal phases produced ondigestion of common formulation lipids and assessment of theirimpact on the apparent solubility of selected poorly water-solubledrugs. J. Pharm. Sci. 2003, 92 (3), 634−648.(23) Kossena, G. A.; Charman, W. N.; Boyd, B. J.; Dunstan, D. E.;Porter, C. J. H. Probing drug solubilization patterns in thegastrointestinal tract after administration of lipid-based deliverysystems: A phase diagram approach. J. Pharm. Sci. 2004, 93 (2),332−348.(24) Harmon, C. M.; Luce, P.; Beth, A. H.; Abumrad, N. A. Labelingof adipocyte membranes by sulfo-N-succinimidyl derivatives of long-chain fatty acids: Inhibition of fatty acid transport. J. Membr. Biol.1991, 121 (3), 261−268.(25) Nieland, T. J. F.; Penman, M.; Dori, L.; Krieger, M.;Kirchhausen, T. Discovery of chemical inhibitors of the selectivetransfer of lipids mediated by the HDL receptor SR-BI. Proc. Natl.Acad. Sci. U.S.A. 2002, 99 (24), 15422−15427.(26) Coort, S. L. M.; Willems, J.; Coumans, W. A.; van der Vusse, G.J.; Bonen, A.; Glatz, J. F. C.; Luiken, J. J. F. P. Sulfo-N-succinimidylesters of long chain fatty acids specifically inhibit fatty acid translocase(FAT/CD36)-mediated cellular fatty acid uptake. Mol. Cell. Biochem.2002, 239 (1), 213−219.(27) Garcia-Calvo, M.; Lisnock, J.; Bull, H. G.; Hawes, B. E.; Burnett,D. A.; Braun, M. P.; Crona, J. H.; Davis, H. R.; Dean, D. C.; Detmers,P. A.; Graziano, M. P.; Hughes, M.; MacIntyre, D. E.; Ogawa, A.;O’Neill, K. A.; Iyer, S. P. N.; Shevell, D. E.; Smith, M. M.; Tang, Y. S.;Makarewicz, A. M.; Ujjainwalla, F.; Altmann, S. W.; Chapman, K. T.;


dx.doi.org/10.1021/mp3006566 | Mol. Pharmaceutics XXXX, XXX, XXX−XXXN

Thornberry, N. A. The target of ezetimibe is Niemann-Pick C1-Like 1(NPC1L1). Proc. Natl. Acad. Sci. U.S.A. 2005, 102 (23), 8132−8137.(28) Staggers, J. E.; Hernell, O.; Stafford, R. J.; Carey, M. C. Physical-chemical behavior of dietary and biliary lipids during intestinaldigestion and absorption. 1. Phase behavior and aggregation states ofmodel lipid systems patterned after aqueous duodenal contents ofhealthy adult human beings. Biochemistry 1990, 29 (8), 2028−2040.(29) Hay, D. W.; Cahalane, M. J.; Timofeyeva, N.; Carey, M. C.Molecular species of lecithins in human gallbladder bile. J. Lipid Res.1993, 34 (5), 759−68.(30) Duane, W. C.; Ginsberg, R. L.; Bennion, L. J. Effects of fastingon bile acid metabolism and biliary lipid composition in man. J. LipidRes. 1976, 17 (3), 211−9.(31) Boyd, B. J.; Porter, C. J. H.; Charman, W. N. Using the polymerpartitioning method to probe the thermodynamic activity of poorlywater-soluble drugs solubilized in model lipid digestion products. J.Pharm. Sci. 2003, 92 (6), 1262−1271.(32) Trevaskis, N. L.; Porter, C. J. H.; Charman, W. N. Bile increasesintestinal lymphatic drug transport in the fasted rat. Pharm. Res. 2005,22 (11), 1863−1870.(33) Caliph, S. M.; Charman, W. N.; Porter, C. J. H. Effect of short-,medium-, and long-chain fatty acid-based vehicles on the absolute oralbioavailability and intestinal lymphatic transport of halofantrine andassessment of mass balance in lymph-cannulated and non-cannulatedrats. J. Pharm. Sci. 2000, 89 (8), 1073−1084.(34) Waynforth, H. B.; Flecknell, P. A. Experimental and surgicaltechnique in rat; Academic Press: 1992.(35) Cummins, C. L.; Salphati, L.; Reid, M. J.; Benet, L. Z. In vivomodulation of intestinal CYP3A metabolism by P-glycoprotein:Studies using the rat single-pass intestinal perfusion model. J.Pharmacol. Exp. Ther. 2003, 305 (1), 306−314.(36) Van Heek, M.; Farley, C.; Compton, D. S.; Hoos, L.; Alton, K.B.; Sybertz, E. J.; Davis, H. R. Comparison of the activity anddisposition of the novel cholesterol absorption inhibitor, SCH58235,and its glucuronide, SCH60663. Br. J. Pharmacol. 2000, 129 (8),1748−1754.(37) Kyle, M. E.; Nakae, D.; Sakaida, I.; Miccadei, S.; Farber, J. L.Endocytosis of superoxide dismutase is required in order for theenzyme to protect hepatocytes from the cytotoxicity of hydrogenperoxide. J. Biol. Chem. 1988, 263 (8), 3784−3789.(38) Winne, D. Rat jejunum perfused in situ: Effect of perfusion rateand intraluminal radius on absorption rate and effective unstirred layerthickness. Naunyn-Schmiedeberg’s Arch. Pharmacol. 1979, 307 (3),265−274.(39) Kossena, G. A.; Charman, W. N.; Boyd, B. J.; Porter, C. J. H. Anovel cubic phase of medium chain lipid origin for the delivery ofpoorly water soluble drugs. J. Controlled Release 2004, 99 (2), 217−229.(40) Johnson, B. M.; Chen, W.; Borchardt, R. T.; Charman, W. N.;Porter, C. J. H. A kinetic evaluation of the absorption, efflux, andmetabolism of verapamil in the autoperfused rat jejunum. J. Pharmacol.Exp. Ther. 2003, 305 (1), 151−158.(41) Wagner, J. G.; Sedman, A. J. Quantitaton of rate ofgastrointestinal and buccal absorption of acidic and basic drugsbased on extraction theory. J. Pharmacokinet. Biopharm. 1973, 1 (1),23−50.(42) Dahan, A.; Miller, J. M.; Hilfinger, J. M.; Yamashita, S.; Yu, L. X.;Lennernas, H.; Amidon, G. L. High-permeability criterion for BCSclassification: Segmental/pH dependent permeability considerations.Mol. Pharmaceutics 2010, 7 (5), 1827−1834.(43) Sugano, K. Estimation of effective intestinal membranepermeability considering bile micelle solubilisation. Int. J. Pharm.2009, 368 (1−2), 116−122.(44) Haikal, Z.; Play, B.; Landrier, J.-F.; Giraud, A.; Ghiringhelli, O.;Lairon, D.; Jourdheuil-Rahmani, D. NPC1L1 and SR-BI are involvedin intestinal cholesterol absorption from small-size lipid donors. Lipids2008, 43 (5), 401−408.(45) Tran, T. T. T.; Poirier, H.; Clement, L.; Nassir, F.; Pelsers, M.M. A. L.; Petit, V.; Degrace, P.; Monnot, M.-C.; Glatz, J. F. C.;

Abumrad, N. A.; Besnard, P.; Niot, I. Luminal lipid regulates CD36levels and downstream signaling to stimulate chylomicron synthesis. J.Biol. Chem. 2011, 286 (28), 25201−25210.(46) Nauli, A. M.; Nassir, F.; Zheng, S.; Yang, Q.; Lo, C. M.;VonLehmden, S. B.; Lee, D.; Jandacek, R. J.; Abumrad, N. A.; Tso, P.CD36 Is important for chylomicron formation and secretion and maymediate cholesterol uptake in the proximal intestine. Gastroenterology2006, 131 (4), 1197−1207.(47) Yu, L.; Bharadwaj, S.; Brown, J. M.; Ma, Y.; Du, W.; Davis, M.A.; Michaely, P.; Liu, P.; Willingham, M. C.; Rudel, L. L. Cholesterol-regulated Translocation of NPC1L1 to the Cell Surface FacilitatesFree Cholesterol Uptake. J. Biol. Chem. 2006, 281 (10), 6616−6624.(48) Moussa, M.; Landrier, J.-F.; Reboul, E.; Ghiringhelli, O.;Comera, C.; Collet, X.; Frohlich, K.; Bohm, V.; Borel, P. Lycopeneabsorption in human intestinal cells and in mice involves ScavengerReceptor Class B Type I but not Niemann-Pick C1-Like 1. J. Nutr.2008, 138 (8), 1432−1436.(49) Reboul, E.; Goncalves, A.; Comera, C.; Bott, R.; Nowicki, M.;Landrier, J.-F.; Jourdheuil-Rahmani, D.; Dufour, C.; Collet, X.; Borel,P. Vitamin D intestinal absorption is not a simple passive diffusion:Evidences for involvement of cholesterol transporters. Mol. Nutr. FoodRes. 2011, 55 (5), 691−702.(50) Reboul, E.; Klein, A.; Bietrix, F.; Gleize, B.; Malezet-Desmoulins,C.; Schneider, M.; Margotat, A.; Lagrost, L.; Collet, X.; Borel, P.Scavenger Receptor Class B Type I (SR-BI) Is involved in vitamin Etransport across the enterocyte. J. Biol. Chem. 2006, 281 (8), 4739−4745.(51) Altmann, S. W.; Davis, H. R., Jr; Yao, X.; Laverty, M.; Compton,D. S.; Zhu, L.-j.; Crona, J. H.; Caplen, M. A.; Hoos, L. M.; Tetzloff, G.;Priestley, T.; Burnett, D. A.; Strader, C. D.; Graziano, M. P. Theidentification of intestinal scavenger receptor class B, type I (SR-BI) byexpression cloning and its role in cholesterol absorption. Biochim.Biophys. Acta, Mol. Cell Biol. Lipids 2002, 1580 (1), 77−93.(52) Goldstein, J. L.; Brown, M. S.; Anderson, R. G. W.; Russell, D.W.; Schneider, W. J. Receptor-mediated endocytosis: Conceptsemerging from the LDL receptor system. Annu. Rev. Cell Biol. 1985,1 (1), 1−39.(53) Mukerjee, P.; Mysels, K. J. Critical micelle concentrations ofaqueous surfactant systems; National Bureau of Standards: Washington,DC, 1971; pp 88−170.(54) Schubert, R.; Schmidt, K. H. Structural changes in vesiclemembranes and mixed micelles of various lipid compositions afterbinding of different bile salts. Biochemistry 1988, 27 (24), 8787−8794.(55) Hofmann, A. F. Bile acids: The good, the bad, and the ugly.Physiology 1999, 14 (1), 24−29.(56) Horter, D.; Dressman, J. B. Influence of physicochemicalproperties on dissolution of drugs in the gastrointestinal tract. Adv.Drug Delivery Rev. 1997, 25 (1), 3−14.(57) Vertzoni, M.; Fotaki, N.; Nicolaides, E.; Reppas, C.; Kostewicz,E. S.; Stippler, E.; Leuner, C.; Dressman, J. Dissolution mediasimulating the intralumenal composition of the small intestine:physiological issues and practical aspects. J. Pharm. Pharmacol. 2004,56 (4), 453−462.(58) Jarvinen, T.; Jarvinen, K.; Schwarting, N.; Stella, V. J. β-Cyclodextrin derivatives, SBE4-β-CD and HP-β-CD, increase the oralbioavailability of cinnarizine in beagle dogs. J. Pharm. Sci. 1995, 84 (3),295−299.(59) Augustijns, P.; Brewster, M. E. Supersaturating drug deliverysystems: Fast is not necessarily good enough. J. Pharm. Sci. 2012, 101(1), 7−9.(60) Tønsberg, H.; Holm, R.; Mu, H.; Boll, J. B.; Jacobsen, J.;Mullertz, A. Effect of bile on the oral absorption of halofantrine inpolyethylene glycol 400 and polysorbate 80 formulations dosed to bileduct cannulated rats. J. Pharm. Pharmacol. 2011, 63 (6), 817−824.(61) Ericzon, B. G.; Todo, S.; Lynch, S.; Kam, I.; Ptachcinski, R. J.;Burckart, G. J.; Van Thiel, D. H.; Starzl, T. E.; Venkataramanan, R.Role of bile and bile salts on cyclosporine absorption in dogs.Transplant Proc. 1987, 19 (1 Part 2), 1248−1249.


dx.doi.org/10.1021/mp3006566 | Mol. Pharmaceutics XXXX, XXX, XXX−XXXO

(62) Behrens, D.; Fricker, R.; Bodoky, A.; Drewe, J.; Harder, F.;Heberer, M. Comparison of cyclosporin A absorption from LCT andMCT solutions following intrajejunal administration in consciousdogs. J. Pharm. Sci. 1996, 85 (6), 666−668.(63) Miller, J. M.; Beig, A.; Carr, R. A.; Webster, G. K.; Dahan, A.The solubility−permeability interplay when using cosolvents forsolubilization: Revising the way we use solubility-enabling formula-tions. Mol. Pharmaceutics 2012, 9 (3), 581−590.(64) Dahan, A.; Beig, A.; Ioffe-Dahan, V.; Agbaria, R.; Miller, J. M.The twofold advantage of the amorphous form as an oral drug deliverypractice for lipophilic compounds: Increased apparent solubility anddrug flux through the intestinal membrane. AAPS J. 2012,DOI: 10.1208/s12248-012-9445-3.(65) Miller, J. M.; Beig, A.; Carr, R. A.; Spence, J. K.; Dahan, A. Awin−win solution in oral delivery of lipophilic drugs: Supersaturationvia amorphous solid dispersions increases apparent solubility withoutsacrifice of intestinal membrane permeability.Mol. Pharmaceutics 2012,9 (7), 2009−2016.


dx.doi.org/10.1021/mp3006566 | Mol. Pharmaceutics XXXX, XXX, XXX−XXXP

Intestinal Bile Secretion Promotes Drug Absorption from Lipid Colloidal Phases via Induction of...

Documents

Transcript of Intestinal Bile Secretion Promotes Drug Absorption from Lipid Colloidal Phases via Induction of...